Transliterated Named Entity Recognition Based on Chinese Word Sketch

One of the unique challenges to Chinese Language Processing is cross-strait named entity recognition. Due to the adoption of different transliteration strategies, foreign name transliterations can vary greatly be-tween PRC and Taiwan. This situation poses a serious problem for NLP tasks: including data mining, translation and information re-trieval. In this paper, we introduce a novel approach to automatic extraction of diver-gent transliterations of foreign named enti-ties by bootstrapping co-occurrence statis-tics from tagged Chinese corpora. In this study, we use Chinese Word Sketch The automatically bootstrapped translitera-tion pairs are further screened based on pho-netic similarity. The precision is evaluated to be more than 90 % against manually cor-rected transliteration pairs.